Skip to main content

Verified by Psychology Today


AI and Optogenetics Disrupt the Neuroscience of Dopamine

Vanderbilt study finds that dopamine is more than just a “pleasure molecule.”

Source: Geralt/Pixabay

Innovative technologies such as artificial intelligence (AI) machine learning and optogenetics are accelerating discoveries in life sciences, especially in the field of neuroscience. A new breakthrough study published in Current Biology by pioneering brain researchers at Vanderbilt University used optogenetics and AI machine learning to reveal that dopamine is not just a “pleasure molecule” — a revolutionary finding that may impact how addiction and psychiatric diseases are treated in the future.

“Dopamine deficits are seen in patients suffering from substance use disorder,” said Erin Calipari, an assistant professor of pharmacology at Vanderbilt University, and faculty member of both the Vanderbilt Brain Institute and the Vanderbilt Center for Addiction Research. “These individuals have reduced dopamine as well as deficits in decision-making that would be explained by our data and new model. These deficits in decision-making are highly correlated with the severity of addiction as well as predicting treatment outcomes. These data are really key to understanding the relationship between dopamine this disease and figuring out how to treat it.”

The neuroscience study was conducted in Calipari’s laboratory under the leadership of both Calipari and postdoctoral fellow Munir Gunes Kutlu. Other members of the Vanderbilt research team include Jennifer Zachry, Patrick Melugin, Stephanie Cajigas, Maxime Chevee, Shannon Kelley, Banu Kutlu, and Cody Siciliano. The Vanderbilt researchers worked in collaboration with Lin Tian, professor and vice chair of biochemistry and molecular medicine at the University of California, Davis.

The existing generally accepted neuroscience theory on dopamine’s role is called the “reward prediction error theory.” This theory posits that dopamine neurons signal a prediction error which is the difference between received and predicted rewards.

A positive prediction error is when dopamine neurons are activated by more reward than predicted, and a negative prediction error is when dopamine neurons show depressed activity with less reward than predicted. Dopamine neurons remain at the baseline activity for fully predicted awards.

Addictive drugs can generate, boost, and takeover the dopamine reward signal and result in uncontrolled dopamine effects on neuronal plasticity. It is theorized that the prediction error message may serve as a powerful signal for behavior and learning.

A shortcoming of the current reward prediction error theory is that it does not describe multiple phenomena linked to dopamine. For example, the reward prediction error theory fails to model dopamine release patterns with negative or neutral stimuli.

In neuroanatomy, the nucleus accumbens is part of the ventral striatum of the basal ganglia. The basal ganglia is a group of subcortical nuclei, or neurons that are located below the cerebral cortex. The nucleus accumbens is the area of the brain associated with reward processing, drug addiction, neurological dysfunction, and psychiatric disorders.

To conduct the experiments, the researchers used optogenetics, a useful technique for neuroscience studies where light can switch genetically modified neurons “on” or “off.” Chrimson was used for recording the dopamine release while stimulating terminals, and channelrhodopsin (ChR2) was used for the optogenetic excitation of dopamine terminals in a separate group of mice.

In the study, the researchers used AI supervised machine learning. Specifically, they used support vector machines (SVMs) classifiers to define whether dopamine located in the nucleus accumbens (NAc) brain area of mice predicted its behavior. They developed a custom MATLAB code to create the training and testing datasets for each dopamine signal linked with behavioral outcomes. The AI algorithm along with a computational model assisted the researchers in analyzing the data.

The researchers demonstrated that dopamine release only conforms to reward prediction error predictions in a subset of learning scenarios, and therefore “pose significant challenges to the hypothesis of dopamine as a prediction error signal while offering an alternative account for the role of NAc core dopamine release as a perceived saliency signal.”

According to the scientists, this study provides a more complete picture of the role dopamine plays in behavioral control and unites multiple theories of dopamine signaling. The impact of this study may have widespread implications on psychopathology—the study of the origin, development and manifestation of abnormal cognition, behavioral dysfunction, and psychological disorders.

“Dopamine is dysregulated in nearly every psychiatric disease such as anxiety, depression, addiction, schizophrenia, and many others,” Calipari said. “To effectively treat patients, we need to know what these dopamine deficits mean and how they regulate behavior. This is the first step to developing new treatments for these disorders.”

Copyright © 2021 Cami Rosso All rights reserved.

More from Cami Rosso
More from Psychology Today
More from Cami Rosso
More from Psychology Today